07-31, 13:45–14:25 (Asia/Taipei), TR309
Language: 漢語
QEMU Fault Tolerance Development Experience - Management and Network Performance Tuning
Abstract –為了讓虛擬機容錯功能可以更方便的使用，我們嘗試將虛擬機容錯技術Cuju (https://github.com/Cuju-ft/Cuju/) 整合進OpenStack雲平台中，本以為只是將底層QEMU虛擬機服務指令傳到上層OpenStack即可，但其實並沒有想像中的那麼簡單，因此本次演講我們將分享這次整合的經驗，包括OpenStack如何和QEMU互動，以及為了和OpenStack整合修改了哪些軟體執行指令以及網路設定方式。在本次演講中，也會分享最近我們最近開發的容錯初始備份最佳化方式以及容錯模式下TCP網路的最佳化，並有影片來DEMO我們將容錯應用於視訊軟體伺服器的情境，以及使用OpenStack開啟容錯功能的展示。
Talk Length –40
您是否知悉並同意如採遠端形式分享，需提供預錄影片（您需同意大會才能接受您的稿件） – yes Difficulty –進階
slido url – hackmd url –https://hackmd.io/@coscup/ryHfQ6vAu/%2F%40coscup%2FBJQ-7awCO
English Abstract –In order to make the virtual machine fault tolerance function more convenient to use, we tried to integrate the virtual machine fault tolerance technology Cuju (https://github.com/Cuju-ft/Cuju/) into the OpenStack cloud platform. We thought it was just the underlying QEMU virtual The machine service command can be passed to the upper layer OpenStack, but it is not as simple as expected. Therefore, in this speech we will share the experience of this integration, including how OpenStack interacts with QEMU, and in order to integrate with OpenStack, which software execution commands and network settings have been modified. In this lecture, we will also share our recently developed optimization methods for fault tolerance initial backup and the optimization of TCP networks in fault tolerance mode. There is also a video to demo the scenario where we apply fault tolerance to video software servers, as well as a demonstration of using OpenStack to turn on fault tolerance.
other info –希望能透過這個演講和同樣是對系統軟體開發有興趣的朋友們一起交流
